I would like to get feedback from people that are more familiar with code base.
I've done a bit of experimenting.
It's not too hard to create a hypercore/hyperdrive in the browser, and synchronize that over a websocket to a gateway, providing you are willing to sync the whole thing. I do this in the dat-shopping-list demo.
It's going to be a little bit trickier if you want to do sparse replication ... together with mafintosh, I was experimenting with hypercore-protocol-proxy which might be a nice way to build a generic proxy.
It would be nice to have a generic Dat protocol proxy. I think rangermauve has done some work in that direction. Unlike IPFS, Dat isn't gateway oriented, but wouldn't it be nice if there was some ability to "bring your own gateway" to any web browser you use (including mobile ones)?
The DatArchive API seems like a logical place to start for an in-browser hyperdrive API, as it's designed for that purpose inside Beaker.
Beaker has support for a library of Dat archives, which is nice because it works across all your webpages -- and it has browser-level support for granting permissions. That would probably require some cleverness to replicate in a conventional browser.